localhoast A record?
cet1 at cam.ac.uk
Fri Mar 21 18:18:54 UTC 2014
On Mar 21 2014, Marco Davids (SIDN) wrote:
>On 21-03-14 14:03, Casey Deccio wrote:
>> I've adopted a number of zones and most of them contain "localhost in
>> a 127.0.0.1" records. I'm curious what current RFC standards state and
>> what the community considers best practice.
>> I would take a look at the query logs for the zones in question. You
>> might be surprised at how many queries are being made by systems that
>> are applying a suffix from the search list because of the lack of of an
>> entry for localhost in the hosts file or the mishandling thereof.
>To me, an NXDOMAIN-reply seems better than an answer with an A-record to
>127.0.0.1 (because that won't be an incentive to fix an apparently
But in the context of search lists an NXDOMAIN will just make the resolver
go on to try the next entry. So in the case of search lists automatically
generated from a "domain" entry, if localhost.astrology.cam.ac.uk doesn't
exist, localhost.cam.ac.uk will be tried, and then localhost.ac.uk ...
>My advice: forget about localhost entries in your zone files, unless it
>concerns a special situation, such as domains that are part of your
Ah, but whose search lists? The resolvers using a particular recursive
nameserver may have many different variants.
>You may want to consider adding it in such a case (although
>I don't do so). But if you do, don't forget to add an AAAA-record for
>::1 as well ;-)
We used to create lots of localhost.[subdomain].cam.ac.uk records, even
to the extent of adding an AAAA record just for those institutions that
had IPv6 enabled on their networks. But we have pretty much given up doing
that for new subdomains. It still seems to me potentially useful to keep
localhost.cam.ac.uk itself, to terminate the probable iteration described
above before it goes any further.
Email: cet1 at cam.ac.uk
More information about the bind-users